When Network Performance Management Systems started out, they largely depended on SNMP and active ping testing to monitor infrastructure uptime and performance. These two data sets allowed them to monitor the up/down status of devices, as well as link usage and errors over time.
This data was great stuff, but it didn’t take long for network problems to outpace this level of visibility.
Before long, when a link would spike with utilization, it wasn’t enough to know that there was a bunch of traffic on a link – we needed to know what that traffic was, and whether it was expected usage or not. To help us, NetFlow entered the scene, which added the “What is that traffic” to the “How Much” in network monitoring.
Active ping tests, SNMP scans, and NetFlow is slick set of data to have, especially when troubleshooting live problems. But today, many application issues have found a way to duck and hide around these visibility methods, evading detection from the tools using them. This creates a situation where the Network Performance Management System may show green lights for network uptime and health, but the application continues to suffer poor performance.